#jogamp @ irc.freenode.net - 20140503 05:05:34 (UTC)
20140503 05:05:34 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20140502050534.html
20140503 05:05:34 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20140503050534.html
20140503 06:18:55 * [Mike] (~Mike]@anon) has joined #jogamp
20140503 07:14:09 * monsieur_max (~maxime@anon) has joined #jogamp
20140503 08:04:20 * zubzub_ (~zubzub@anon) Quit (*.net *.split)
20140503 08:04:20 * magaio (~magaio@anon) Quit (*.net *.split)
20140503 08:06:51 * zubzub_ (~zubzub@anon) has joined #jogamp
20140503 08:06:51 * magaio (~magaio@anon) has joined #jogamp
20140503 16:35:10 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20140503 16:39:58 * hharrison (~chatzilla@anon) has joined #jogamp
20140503 16:40:26 <hharrison> Java3d 1.6.0-pre10 released...for those that still care ;-)
20140503 19:35:39 * hija (~hija@anon) has joined #jogamp
20140503 21:01:14 <zubzub_> according to my ide gluegen-rt-main 2.1.5-01 pulls ine jogl-all-main 2.0.2-rc2-20130404
20140503 21:01:26 <zubzub_> which is a problem because
20140503 21:01:30 <zubzub_> it shouldn't do that
20140503 21:01:37 <rmk0> it certainly shouldn't
20140503 21:01:39 <zubzub_> I want to use jogl-all-noawt-main
20140503 21:01:57 <rmk0> depressing... let me see what's gone wrong this time
20140503 21:02:02 <zubzub_> w8 just a sec
20140503 21:02:14 <zubzub_> let me verify first that iit's not nifty gui doing it
20140503 21:03:05 <rmk0> ... just realized that jogl-all-main doesn't depend on gluegen
20140503 21:03:12 <rmk0> forgot we kept them separate
20140503 21:03:18 <zubzub_> idd
20140503 21:03:27 <rmk0> is likely another transitive dependency, as you suspected
20140503 21:03:32 <zubzub_> 'looks like it's idd nify
20140503 21:03:39 <zubzub_> and my ide giving false reporting
20140503 21:03:54 * zubzub_ submits bug report to nifty
20140503 21:04:29 <rmk0> nifty's actually depending on artifacts that are free to vanish at any moment
20140503 21:04:45 <rmk0> no rc ever goes on maven central
20140503 21:05:03 <rmk0> i did put that on the wiki... maybe i should use a bigger font and blink tags
20140503 21:10:09 <zubzub_> nifty needs to make it's jogl dependencies 'provided' instead of 'compile' so they're not transitive
20140503 21:10:45 <zubzub_> else I'll have to explicitly override *all* resolved jogl deps
20140503 21:11:34 <rmk0> it can't really do that safely unless we honour api versioning
20140503 21:11:44 * rmk0 throws rotten fruit at sgothel
20140503 21:11:46 * rmk0 hides
20140503 21:12:24 <rmk0> nifty can say that it wants [2.1.0, 2.2.0) in provided scope
20140503 21:12:46 <rmk0> but... 2.1.5 broke compatibility with 2.1.4 without incrementing the minor version
20140503 21:13:14 <rmk0> kind of fucks things up
20140503 21:16:43 <zubzub_> semantic versioning!
20140503 21:17:04 <rmk0> i wish
20140503 21:17:33 <zubzub_> reminds me of the problems at work as a couple of self proclaimed exports made a complete mess of versioning our modules
20140503 21:17:39 <zubzub_> *experts
20140503 21:17:54 <rmk0> is not even difficult
20140503 21:18:10 <zubzub_> yup
20140503 21:18:19 <zubzub_> osgi even has built in support for it
20140503 21:19:49 <rmk0> been eyeing osgi a bit lately
20140503 21:19:55 <rmk0> don't really know anything about it
20140503 21:20:17 <rmk0> got no use for the "application server" side of it, but the packaging side looks somewhat better than maven's
20140503 21:20:30 <zubzub_> it's not really comparable to maven
20140503 21:20:37 <zubzub_> they serve and solve different problems
20140503 21:20:44 <rmk0> just meant the conventions they seem to have for packages
20140503 21:21:10 <rmk0> have you used it in any capacity?
20140503 21:21:14 <rmk0> not... actually met anyone that has
20140503 21:21:25 <zubzub_> I use it at work on 20000 devices
20140503 21:21:33 <rmk0> hehe, right
20140503 21:21:42 <rmk0> think that probably counts as "some capacity"
20140503 21:21:42 <zubzub_> :)
20140503 21:21:53 <zubzub_> public transport sector
20140503 21:21:58 <zubzub_> payment systems
20140503 21:22:15 <zubzub_> high end embedded devices really
20140503 21:22:24 <rmk0> am wondering... does it allow multiple versions of a package to exist in the same VM?
20140503 21:22:29 <zubzub_> yup
20140503 21:22:33 <rmk0> ceylon use custom classloaders for that, assume it does the same
20140503 21:22:36 <rmk0> *uses
20140503 21:22:40 <zubzub_> yep
20140503 21:22:43 <rmk0> they use one classloader per package
20140503 21:22:49 <zubzub_> same with osgi
20140503 21:23:03 <rmk0> it seems like a vastly better way to do things
20140503 21:23:08 <zubzub_> as osgi exists from 1998 I guess they 'stole' it from osgi ;)
20140503 21:23:21 <rmk0> i think they stole it from jboss modules, who probably stole it from osgi
20140503 21:23:29 <zubzub_> that's whaty really makes osgi powerfull
20140503 21:23:52 <zubzub_> as it does dependency matching/resolving at startup
20140503 21:23:59 <zubzub_> of different jars (bundles)
20140503 21:24:04 <zubzub_> by looking at their manifests
20140503 21:24:09 <rmk0> right
20140503 21:24:27 <zubzub_> so if it encounters a jar that can not be resolved, it will not even be touched
20140503 21:24:50 <zubzub_> or you can deploy 2 jars of the same product with 2 versions
20140503 21:25:09 <zubzub_> and osgi will use the semantic versioning to match it to other jars depending on it
20140503 21:25:35 <rmk0> ♥
20140503 21:25:57 <zubzub_> but with great powre comes great responsibility
20140503 21:26:15 <zubzub_> as classleaders are now more a bit private
20140503 21:26:28 <zubzub_> so you cant simply go and start loading classes from other jars
20140503 21:26:47 <zubzub_> classForName is a nono
20140503 21:26:52 <zubzub_> (wont work in osgi)
20140503 21:26:54 <rmk0> sounds good to me
20140503 21:26:59 <zubzub_> and there's a lot of code doing it
20140503 21:27:10 <zubzub_> same for classpath scanning & loading
20140503 21:27:23 <zubzub_> jpa for example
20140503 21:27:43 <zubzub_> luckly they added a special jpa api to work aroudn the problem
20140503 21:28:04 <zubzub_> anyway I'm going to bed
20140503 21:28:06 <zubzub_> nn
20140503 21:28:08 * rmk0 waves
20140503 23:04:07 * magaio (~magaio@anon) Quit (Read error: Connection reset by peer)
20140503 23:06:58 * magaio (~magaio@anon) has joined #jogamp
20140504 01:05:39 * hija (~hija@anon) Quit (Quit: hija)
20140504 01:09:51 * hija (~hija@anon) has joined #jogamp
20140504 01:25:20 * hija (~hija@anon) Quit (Quit: hija)
20140504 05:05:34 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20140504050534.html